One benefit of semi-retirement has been reconnecting and brainstorming with old customers, employees, and peers without the management and vendor partner requirements from running a software and consulting company.  My firm, Technology Services Group (TSG, bought by Alfresco/Hyland in 2020), specialized in content management for over 24 years and I worked in the space for 6 years before that for Andersen Consulting (now Accenture).  Since leaving Alfresco/Hyland back in 2020, several old business connections have told me, “Dave – we miss your ‘What’s next with ECM’ posts.”  This post will share my thoughts on the history of the industry and predictions on the potential future of the industry along with some thoughts on the vendors with my now complete independence from my old ties to any of the ECM vendors.

History of Enterprise Content Management

Whether we called our niche Image and Document Management (IDM, mid 1990s), Enterprise Content Management (ECM, late 1990s) or Content Services Platform (CSP, 2016ish and beyond), the major vendors in the space surprisingly haven’t really changed over the years compared to other technologies (e.g. databases, phones, or browsers).  Predicting where the industry is going in 2024 requires understanding where the industry has been. 

One of the industry’s more notable predictive posts was back in 2017 by Gartner Analyst Michael Woodbridge, where in the Gartner blog he proclaimed “The Death of ECM and Birth of Content Services.”  The original push for Enterprise Content Management in the 1990s followed the SAP moniker of Enterprise Resource Planning (ERP), where initially SAP could easily justify the efficiency of one ERP software platform for the entire company rather than tying multiple different accounting systems together.  ECM pushed for that same focus and many clients built multiple solutions around one ECM infrastructure.  However, the hope for “one repository/vendor to rule them all” failed the ECM industry for multiple reasons, including:

• Hardware, Database and Storage Efficiencies – Original late 1990s approaches all looked to leverage one infrastructure, and often server rack, for all the content management components for efficiency and cost.  I have recollection of installing designated ECM databases and content storage RAID arrays for multiple clients.  With the evolution to Storage Area Networks, Network Computing, Cloud Computing, and Solr/Elasticsearch, more of the infrastructure can be easily and cost-effectively spread across the network removing the cost savings for one dedicated shared content management solution.
• Fault Tolerance and Security – One of the hopes for an organization having a single ECM infrastructure was to reduce the likelihood of a failure and consolidate support resources.  Unfortunately, the more groups that shared an ECM infrastructure, the more likely there could be a risk to faults or security, either because of one specific group or through complexities in managing the groups’ documents together.  We had one pharmaceutical client that went through a huge effort to consolidate Documentum systems with R&D and Manufacturing, only to have a critical fault from R&D bring both systems down.  The client unwound the consolidation effort—at considerable expense—to separate the systems.
• Failure of the “Enterprise” Search Scenario – ECM enthusiasts had a vision that typically was vocalized as “Wouldn’t it be great if you could access all the documents of the enterprise from one interface?”  Similar to the failures of Federated Search that I blogged about in 2019, Enterprise Search is a concept that sounds workable but is difficult to implement due to security, complexity, and costs, and is a solution that’s rarely ever actually needed by users to justify the expense in creating it.
• Failure of the ECM Vendors – Unlike Storage Area Network and cloud provider networks where the complexity and cost per document has dropped dramatically over the last 20 years, the ECM vendors have not reduced the costs of their offerings or simplified the database and infrastructure components.  Often the support has gotten more complicated given the ECM suite approach and older technology.  The costs per user and document have also increased despite the infrastructure costs being reduced.  See the post I wrote back in 2016: “Is a Large ECM Suite Really that Sweet?”

Most current ECM infrastructures are for one application, and therefore not truly enterprise systems.  Most large customers have five or more content-related vendor solutions.  Given the evolution of the market, some commonalities of the old ECM vendors include:

• ECM Feature Creep and Bloat – One issue that stubbornly remains from the enterprise approach days is “feature creep,” where vendors were selected based on number of features.  Back in the day, vendors competed on ECM RFPs with a “feature shoot out” competition.  Vendors built rarely-used capabilities into their core system to win the feature competition and the RFP process.  Rarely-used ECM features like document level security, virtual documents, and branching versions were added to the core database schema, APIs, and interfaces, which resulted in bloat and frequent scaling and support issues.
• “Cash Cow” Status – Following the BCG Matrix, old ECM Vendors are in the mode of low growth with high market share, where they look to milk cash from their long term customers.  Slow market growth as well as difficulties in dislodging existing ECM vendors discourages new innovators from directly challenging the existing vendors.
• Support and Service Personal Difficulties – Given the talent push for other newer technologies and cost-cutting for improved profits, vendors have struggled to attract, retain, incent, or motivate existing or new technical resources.  Customers are frustrated with high costs for low-quality levels of support or service.
• Developing and Selling Add-ons Rather than Core Innovation – With an entrenched customer base of thousands or millions of documents in the document repository, growth or additional sales are typically add-ons rather than disruptive or innovative changes to its core.  Core changes would require migration that might cause customers to think of moving off the platform and stop paying the vendor.  Over the last 6 years, most of these add-ons include movement to vendor-supplied cloud infrastructure, but I would anticipate that other add-ons would follow market trends.  Given the current hype around AI in 2023, I would expect vendors to offer additional add-ons focused on AI.

Review of Previous “What’s Next” Post

My last post on an ECM Roadmap from 2017 is available on the TSG Blog site.  A recap:

• Challenges for any roadmap included budget, resources, as well as the “kick the can down the road” delay of moving platforms.  The resource issue had gotten even worse as ECM consultants and engineers moved on to other industries and technologies without the vendors able to hire and train replacements.
• When it came to the different vendors, at the time even TSG stated our biases toward Alfresco (note – Alfresco acquired TSG in 2020).  IBM was an early adopter of AI and very focused on “Watson” (if you remember all those ads, you might have noticed they’re coming back) and their solutions (FileNet, CMOD, DB2 Content Manager…) were all in “Cash Cow” or minimal maintenance mode.  OpenText had just acquired Documentum and there was concern about whether OpenText would maintain the Documentum platform or try to merge the customers into some new or old OpenText platform.  We saw Hyland as a niche in the healthcare space and I incorrectly spelled Hyland as “Highland,” showing how little we knew of the company and their space around health care (as well as my need for better editing/review).
• My predictions for ECM disruptors were infrastructure as a service, browser-based viewing, software as a service, and open source.

Since 2017, the infrastructure and software as a service have come to fruition with more and more clients moving to the cloud, either on their own or as part of a vendor offering.  Browser-based viewing has evolved and is common.  Open Source has changed the most dramatically with industry leaders Alfresco and Nuxeo being bought by Hyland.

Understanding the Vendors in 2023

In taking a fresh look at OpenText with the Documentum acquisition in 2017, as well as Hyland with the acquisition of Alfresco in 2020 and Nuxeo in 2021, I would propose that OpenText and Hyland have followed the roadmap established by IBM back when they acquired FileNet in 2006.  After any acquisition, the acquirer needs to justify the purchase with profits to pay back the cost of the acquisition.  Increasing profits involve either increasing revenue or decreasing costs.  Similarities in the three vendors post-acquisition include:

• Staff Reduction – One example that many technology firms are following is that of X, formerly Twitter, where Elon Musk seems to be running the platform with 80% less people.  Whether moving jobs off-shore or consolidating multiple departments (think: marketing), all the firms have decreased costs and increased profits by reducing headcount, just as Hyland did in 2023 with a 1,000 person layoff. This only exacerbates the reduction in the quality of support and services.  Similar layoffs and attrition have happened at IBM and OpenText.
• Keeping Existing Customers Paying for Lower Quality Services – The subtle part of cost reductions for software firms is keeping their existing customers paying the same in maintenance or subscriptions despite staff reductions affecting the quality of the customer’s experience.  An example that I often use is considering cost reductions for a pizza company.  Reducing costs is as simple as putting less cheese or other ingredients on the pizza.  The problem for pizza companies (and software firms) is just how much cheese can you remove before the customer notices that the pizza is not good anymore?  One advantage all the vendors have is that it is difficult (and expensive) to change platforms, and all vendors are not incented to make migration easy.
• Get New Customers from Other Vendors – While the ECM niche is not a growing marketplace, vendors can still get new customers by moving customers off older platforms.  TSG was very successful moving IBM/FileNet customers to Alfresco.  I would anticipate that both Hyland and OpenText will continue to pursue other vendor’s install base, but those opportunities are dwindling.
• Get Existing Customers to Pay More – One strategy the vendors use successfully is either increasing the cost of the customer’s current environment or selling additional components, capabilities or services.  Every vendor is taking this approach with addons of cloud, increased license/maintenance, services, or other ways to have existing customers pay more.  Alfresco, after acquiring TSG, was very successful in selling the addons of the Alfresco Enterprise Viewer and Alfresco Content Accelerator to both new and existing customers as add-ons.  We would anticipate that add-ons like AI will be coming as vendors look to stay relevant as well as retain their customer base from competitors.    Another approach includes increasing license/maintenance cost.  As another way to get clients to pay more for existing installations, we saw Documentum increase both maintenance from 18% to 27% as well as increase aggressive client audits.  Lastly, legacy firms have always offered consulting services as another way to get customers to pay more.  Unfortunately, software firm consulting services rarely satisfy customers’ needs.  Competing for consulting services also makes legacy software firm less partner-friendly and reduces the ecosystem of knowledgeable resources available for the customer.  I blogged about the ups and downs of partnerships at Founder-In-Eleven with details from the TSG’s Documentum partnership.

If Cash Cow vendors are not delivering the value for the cost of the software and services, innovative customers must develop a roadmap that gives them the flexibility to either stay with the existing vendor if the value is good or reduce the costs of the vendor to correspond with the value provided.

Documentum, FileNet, Hyland, and OpenText – “Should I Stay or Should I Go?”

In 2017 I wrote an article for Alfresco regarding the acquisition of Documentum by OpenText: “Documentum, FileNet & OpenText – Should I Stay or Should I Go?”  I stated that IBM, 10 years after the FileNet Acquisition, was focused on consulting, a Proprietary IBM Cloud, Box and Watson as opposed to improving FileNet.  At the time I surmised that OpenText would follow a similar roadmap of IBM and not invest in Documentum:

“It seems OpenText is more focused on the “what’s next” in regards to analytics, robotics and their proprietary cloud rather than how Documentum (or OpenText Content Manager) would be improved.”

With cloud a focus, I surmised that clients would be more successful looking at Amazon or Microsoft for cloud infrastructure rather than proprietary IBM or OpenText clouds.  Alfresco had developed their cloud offering on Amazon and the article presented those benefits versus the proprietary clouds.  That summary, which incorporated another The Clash reference of “Should I cool it or should I blow,” was:

“With the cloud, cost savings and rapid growth of Infrastructure as a Service (IAAS) vendors like AWS, legacy ECM customers are considering cloud ECM alternatives. While concern about the effort of moving documents and integrations to a new platform has in the past pushed clients to delay leaving, the cost savings and capabilities of AWS combined with Alfresco can dramatically reduce costs and make the “should I go” dilemma that much easier.”

With a fresh 2023 take, the Alfresco cloud has been very successful and my understanding is that it has become a strong cloud offering from Hyland.  One thing for clients to consider is that the Alfresco cloud offering is not an “Infrastructure as a Service” (IaaS) offering that customers install in their own version of Amazon, but a “Platform as a Service” (PaaS) model where the customers are on a proprietary Hyland offering built on Amazon.  While this approach is better than the previous Hyland, IBM, or OpenText approaches built on non-Amazon infrastructures, the Alfresco PaaS cloud does share the proprietary and lock-in components that can result in pricing issues and difficulties to move from the Alfresco cloud.

ECM Futures – The Innovation Dilemma for Hyland, OpenText, and IBM

One of the books I always recommend for software strategy consideration is The Innovator’s Dilemma by Clayton M. Christensen.  In talking about technologies, the book highlights the difference between disruptive and sustaining technologies.

• Sustaining technologies focus on growing existing technologies by enhancing their performance, mostly through extended functionality or increased capacity.
• Disrupting technologies change the landscape of the industry, or spark a new industry altogether, because they solve a problem in an entirely new way or for an entirely new group of people.

In looking at disruption for the ECM vendors, one of the best examples with a document focus is how Google Docs disrupted Microsoft Office.  Microsoft Office was sustained over time by adding more and more features without disrupting the core requirement of needing Office to work on documents.  Google Docs is disrupting with both the technology of browser-based document editing, price (free), as well as the collaboration of editing a document with other users at the same time.  While Google Docs has not replaced Microsoft Office, it has disrupted its use while “nibbling” away at former Microsoft Office customers.  For example, many school districts have standardized on Google Docs, Google Classroom and Chromebooks based on cost and the technology.

Hyland, OpenText, and IBM have been very fortunate that most of the technology innovation over the past 25 years has been more sustaining than disruptive.  Whether SAN/Object Store, Database, Network, Browser, or Solr/Elasticsearch, these technologies have not disrupted the core repository and function of the ECM system.  However, there have been some notable new entrants in the space over the last 20 years:

• Box – Playing off the cloud and software as a service, Box has been successful in introducing content outside the organization, which initially started as a “Freemium” business model.  Box did pivot to include ECM back in 2014 after initially saying they were not interested in pursuing those specific scenarios.
• Alfresco and Nuxeo – Both started with open source to provide lower-cost disruptive technologies.  Both of these offerings have since pivoted to more commercial, higher-cost software and have been purchased by Hyland in 2020 and 2021 respectively.
• Veeva – Leveraged their success with Salesforce and Software as a Service to offer content management for life sciences, and has been very successful in moving former Documentum customers to Veeva.
• AODocs – Leveraged Google infrastructure for technology disruption.
• NextDocs – Leveraged Microsoft SharePoint infrastructure for technology disruption.

Each of the above has been successful in “nibbling” at the install base of the older vendors, but unlike Google Docs, has not been able to truly disrupt the current ECM vendors with lower cost or substantially better technology.  With the market in Cash Cow low-growth status, it is difficult for any new vendors to disrupt without compelling technology or cost-reduction capabilities.

The Coming Disruption of AI

Legacy ECM vendors will treat AI as the next sustaining technology for their products.  Similar to machine learning in 2014, look for vendors to leverage AI by automating many of the existing interface capabilities of their systems and repositories.  Vendors will promote new AI capabilities including content classification and tagging, advanced search, content recommendations, automated workflows, intelligent document capture, and content summarization, all of which represent an evolution toward machine-learning goals. 

I would predict that, rather than a sustaining technology for the ECM marketplace, AI could be a substantial disruptor as it offers the ability to change the interface itself.  Traditional content interfaces focus on “get me the documents I need.”  AI changes the context to more of a “solve this problem with the data you have been trained on.”  To quote Bill Gates in his 2023 article “The Age of AI Has Begun,” he says, “I knew I had just seen the most important advance in technology since the graphical user interface.”  Here, Gates certainly wasn’t talking about tagging documents.  By changing the user experience with smarter, personalized, conversational and natural language interfaces, AI disrupts the need for a document search and retrieval interface for many business scenarios.

Customers incorporating AI for their own internal use will leverage large language models developed by a variety of vendors from publicly available text data.  These will include the major cloud offerings including Amazon, Microsoft, and Google. After starting with a commercial large language model, customers will want to fine-tune the model on more specific datasets, including proprietary or internal documents that can allow the model to adapt to the specific language and domain of interest.  The fine-tuning process involves training the model on a narrower dataset with labeled examples and specific tasks relevant to the intended application. For instance, if the goal is to assist with customer support queries, the model might be fine-tuned on a dataset of customer support interactions.

As the fine-tuning involves iteration and trial and error, I would predict that the cloud offering  tools will require the customer’s specific datasets to be stored and properly tagged in their own cloud-based repositories.  These cloud distributors will not support the current ECM vendors’ cloud instances, repositories, or other integration to legacy databases, internal systems, and file systems. 

Innovative customers looking to pursue AI will either develop an on-going migration approach to move training content to the cloud repository, or begin managing their content directly in the large cloud providers repository.

The Path Forward – A Clean Slate Approach

With AI and cost-savings in mind, innovative ECM vendors and clients are starting to take a “clean slate” look at their content and document requirements along with what technology and interfaces are necessary.  A clean slate approach allows for the removal of both ECM RFP feature creep and old school interfaces and technologies.  My predictions of what that future state could look like include:

1. Built from the Object Store – Object Stores like AWS S3 are disrupting ECM tools with capabilities like attributes, security, tremendous scale, and lower costs.  Innovative vendors and clients can build out what were typically ECM scenarios without any ECM vendor cost or technologies.  Addition of web-based capacities that allow content to be accessed via RESTful APIs over HTTPS allows for easy integration to cloud-native applications. See Docuvela’s excellent post comparing the traditional content store to the cloud object store.  Once in the cloud object store, content can be easily leveraged for fine-tuning AI models.
2. Purpose-Built Interfaces with Modern Tools – More vendors and clients will build configurable interfaces focused on business scenarios with modern, cloud-native microservices architectures and AI rather than generic “old school” ECM feature creep file/cabinet approaches.  Modern tools provide for cost reductions, better scale, and the ability to attract and retain internal development resources who will not want to work or support old ECM interfaces.  See Docuvela’s post comparing scalable microservices to traditional three-tier approaches.
3. Built without a Relational Database – For matters that require complex content scenarios, vendors and clients will look to modern tools like Solr/Elasticsearch/NoSQL for storage and retrieval of content related data.  The relational database behind traditional ECM vendors has always been a source of frustration for clients in regard to cost and support, as most support issues always seem to be a result of database issues (e.g.  table space, corrupted indices, or storage allocation).  Modern tools remove both the cost and support issues while providing superior scale and performance as well as prepare for fine-tuning AI models.
4. Built with a Consulting Partner – The resource issues for IT departments to hire and retain developers have only gotten worse over the last 20 years.  Large non-technology firms will not be able to attract and retain experienced ECM, Amazon, or Microsoft resources, and instead will partner with consulting firms that bring knowledge to build and support new efforts while helping to maintain legacy ECM systems.

Typically, after any prediction or opinion posts, at least one ECM vendor will respond with “We are already doing that.”  See my back and forth with different IBM resources on LinkedIn back in 2017 when I posted on “FileNet – Will IBM Sell It?”  While IBM didn’t sell FileNet (I predicted they wouldn’t), IBM never really released any of the product improvements their resources were so passionate in defending.  As it relates to the approach above, here are my reasons why the existing vendors will be disrupted rather than turning the technologies into sustaining innovations for their products:

1.  Cost – With a market in “Cash Cow” Status, existing vendors will choose not to compete with themselves by offering new technology at lower costs.  All the ECM vendors are in a “keep the customer paying the same or more” mode.  Without a lower cost offered by the vendor, the customer has no incentive to take the risk and pay the cost to move.  Customers will only take the risk to move to new vendors that offer lower costs and lower support risk.
2.  Perception – New disrupting technologies make old technologies look stale, which can disrupt existing sales pipelines of current products and the products’ support engineers.  I experienced this concern at Alfresco with TSG’s NoSQL alternatives to Alfresco’s traditional bloated ECM relational database.  Despite touting TSG success with massive scale in the press release announcing the purchase of TSG, Alfresco never released NoSQL or allowed us to promote it as an alternative.  As a current example, Hyland’s NoSQL/MongoDB alternative from Nuxeo is also something that is rarely promoted.
3. Internal Conflict and Ability – Disruption in the technology space requires both the knowledgeable resources and the ability to move fast and make mistakes.  For those familiar with the Apple journey, the conflict that arose between the Apple II and Macintosh teams is a cautionary tale of that drama.  Existing vendors that are public or looking to be sold will not seek the disruption drama and do not have the resources or stamina to truly pivot to disruptive business models.

Conclusion

With the main vendors currently in “Cash Cow” mode, I predict that the existing vendors will build out solutions that sustain their repositories using AI, workflow, or new add-on components that increase customer cost rather than disrupt their existing infrastructures or lower prices. 

Innovative customers will look to build more capabilities to move content to cloud object stores for cost savings as well as AI fine-tuning through either migrating/copying the content from legacy vendor solutions or managing the content directly in the object store.

Given the lack of resources in the ECM space, and the difficulty of customers to train or retain employees with robust ECM legacy vendor experience, customers will need access to a consulting partner that not only has the skills surrounding older technologies to support their existing infrastructure, but also the vision to help clients realize a “nibble” roadmap that moves toward cloud-based object stores and AI. Innovative consulting firms like Docuvela will develop solutions that augment their consulting services with starter software for integration with Amazon, Azure, and other cloud-based object stores in 2024 in preparation for the upcoming AI disruption.